RESUMEN
AIMS: To assess institutional compliance with, and test characteristics of, a child abuse screen performed by emergency department (ED) nurses for children <5 years old who were diagnosed with fractures. METHODS: A secondary analysis of a retrospective observational study of children 0-5 years old with fractures seen at a pediatric ED between January 2018 and April 2023 was performed. We analyzed demographics, ED visit data, and results of the nurse-completed abuse screen. Screen results were compared to ED clinician concern for abuse to calculate test characteristics. RESULTS: The mean age of the 2,705 children identified was 38.4 months (SD 19.8). Out of the total patient cohort, 2,449 (90.5%) had a nurse-completed screen. Among these, 65 patients (2.4%) screened positive for possible abuse. We found no statistically significant difference in screen completion by age group, race, ethnicity, language, or insurance type. Of 312 (11.5%) encounters with clinician concern for abuse, 17.6% screened positive, 76.0% screened negative, and 6.4% had an incomplete screen. The sensitivity and specificity among screened children aged 0-5 were 19.2% [95% CI 14.7-23.8%] and 99.5% [95% CI 99.3-99.8%]. The PPV and NPV were 84.6% [95% CI 75.8-93.4%] and 90.3% [95% CI 89.1-91.5%]. Comparatively, among children <12 months, the sensitivity was 24.4% [95% CI 18.0-30.8%], specificity was 98.1% [95% CI 95.4-100%], PPV was 95.5% [95% CI 89.3-100%], and NPV was 43.7% [95% CI 37.3-50.1%]. CONCLUSIONS: Although there was high compliance with this nurse-completed abuse screen, it is an inadequate sole modality for screening young children with fractures, with a low probability of a positive screen given clinician concern for potential abuse for the entire cohort and among high-risk infants.